Edition 2026.2 — main residence in the existing market

Can you still become a homeowner in France?

Price, rates, incomes, deposit and resale: what has really changed purchasing capacity in property, and what the figures published on 15 September 2026 allow us to test.

The answer

The same monthly effort finances less capital than at the rate trough, while the price of existing homes remains above its 2000 level: it is the combination of price and rate, and not the price alone, that decides, and the latest published readings do not ease it over one year.

At a fixed monthly payment and term, the capital a bank can lend is the inverse of a function of the rate: without the household changing its budget, what it can borrow moves in the opposite direction to the rate.

What this does not say

What this study does not settle: whether to buy, wait or give up. It carries no local price, no buyer-household income, no cost of use — no avoided rent, no works, no property tax. It shows a mechanism and gives what is needed to redo it; the decision remains entirely open.

Where do we stand? The latest published readings

At the edition of 15 September 2026: the latest rate month and the latest price quarter published by their producers.

Rate for the latest published month (July 2026): 3.18%
Existing-home prices, second quarter of 2026, over one year: −0.8%
What the same monthly payment buys, over one year: −0.8%

Prices are falling, and it has not become more affordable: over one year, the same monthly payment buys less than a year earlier. A provisoire quarter of prices and one month of rates do not make a trend; they say where we stand on the day this study appears.

Reserved · full analysis

The quantified, dated answer, over 26 years of rates and 30 years of prices, extended to the latest published rate month — and the state of the last closed year, benchmark by benchmark

Open to all

Three situations, followed from end to end

The study tells nobody's story. It sets out three CALCULATION SCENARIOS, whose income derives from the median net wage published by INSEE multiplied by a declared number of wages, and whose monthly burden is 30% of that income — below the debt ceiling of 35%. The same three situations then run through all chapters, so that a result is always compared with the same starting point.

Situation A
Single income, first purchase
Income used
€2,190 a month
one median net wage
Credit burden
€657 a month
Deposit
10% of the price
What this scenario is not: a single median income, and a deposit that only just covers acquisition costs: only a marginal share is left to reduce the amount borrowed. The scenario is therefore the most exposed of the three to a fall in prices.
Situation B
Two incomes, first purchase
Income used
€4,380 a month
2 median net wages
Credit burden
€1,314 a month
Deposit
20% of the price
What this scenario is not: two median salaries do not make a median household: the income distribution of buyer couples is not that of employees.
Situation C
Two incomes, already a homeowner
Income used
€4,380 a month
2 median net wages
Credit burden
€1,314 a month
Deposit
35% of the price
What this scenario is not: the higher deposit represents what a previous sale would have left the household; this scenario ASSUMES it, whereas the chapter on the second purchase calculates it from an actual sale in scenario B.

What this does not say

These three situations are not observed households. Buyers form a particular subset of the population, for which the data catalogue carries neither income, nor deposit, nor age. None of the three describes “the average French person”, who exists in no series.

Reserved

What one point of rate really changes

How much does one point of rate remove from financeable capital, and why is the effect not the same depending on the level from which one starts?

Reserved · full analysis

What one point of rate really removes from financeable capital, over 2 terms and five rate levels, step by step — with the reason why the effect is not constant

Reserved

How much did the same monthly effort allow one to borrow, depending on the period?

At identical monthly payment and term, how much could a household borrow depending on the year of signing?

Reserved · full analysis

What the same monthly effort allowed one to borrow depending on the period, year by year from 2000 to 2025, at the rate actually observed — the most favourable year, the least favourable and the last closed year

The plots

What the series show

The price of existing homes since 1996 — a constant-quality index — open to all. In the full study, financeable capital against the rate and what is still owed in two market environments.

1 free chart2 premium chartsThe premium panels work — they are deliberately closed. Click one to see what it contains.

The price of existing homes since 1996

By how much have existing-home prices moved, and since when?

FreeObserved 1996-2025Index, base 100 in the 2015 annual average

Reading

All dwellings
Apartments
Houses

Source — INSEE — Notaires-INSEE price indices for existing homes, metropolitan France. Axis: the axis does not start from zero: a constant-quality index is read through its relative changes, and an axis starting from zero would squash a movement of more than two hundred per cent into a single slope

What to look for

The overall index goes from 41.5 in 1996 to 126.9 in 2025.
The highest point of the series is reached in 2022, at 132.4.
From 2022 to 2025, the index fell by 4.2%.
Apartments and houses do not follow exactly the same slope: the gap widens before 2008, then closes.

Why it matters

This is the backdrop to everything else: what a home costs decides what has to be borrowed, and therefore what the rate does to the project.

Reading of the selected year

Hover over the chart to read a year in the tooltip, click to hold it here — or use the keyboard arrows.

What this does not mean

This is not a price: the index carries neither euros nor square metres.
This is not a local market: the dispersion between municipalities far exceeds the national trend.
This is not new housing, which follows other prices and other transfer costs.

What this chart cannot show, and how to use it

Limits

The index is constant-quality: it neutralises the fact that the homes sold in one year are not those sold the next.
The series is raw, not seasonally adjusted; the year is the average of its four quarters.

Interactions

Hover over the plot: a tooltip gives the point's values, unit and status. Click to hold the selection — it does not disappear when the pointer moves away, so that the reading stays displayed while you read it. On the keyboard, the tab key brings focus to the plot, the arrow keys move the selection, Home and End go to the ends, and Escape closes. On a touch screen, one press selects a point, a second releases it.

Open to all

What is said, and what the figures make of it

9 claims in circulation, set against the 20 frozen conclusions of this study. 2 others were set aside, and the reason is written further down.

The verdict is not written: it is calculated. Each conclusion of the study is declared as going in the direction of the claim, against it, or relevant without settling — and a single rule turns these positions into a verdict. Conclusions on both sides give “misleading”; no conclusion that settles gives “cannot be settled”.

Rather true · 1
Misleading · 4
False · 1
Cannot be settled · 3

Misleading
“You have to keep a home for at least five years so as not to lose money.”
The detailed answer, and the measurement that underpins it, can be read in the full study.
What this verdict does not cover: the thresholds measure no profitability: they ignore the rent that would have been paid without the purchase, the return the deposit would have produced elsewhere, works and property tax. “Losing money” in the full wealth sense is not what they calculate.
See why— 2 conclusions of the study
Why this idea circulates: the rule circulates because it blends two true things: amortisation does reduce what the sale has to cover, and acquisition costs are indeed an immediate cost. From them a single period is drawn, whereas the two movements do not follow the same calendar.
- supports it · value withheld
  The study quantifies, after eight years of amortisation, the price fall tolerated by the cash threshold and by the deposit-recovery threshold.
- goes against it · value withheld
  The study quantifies the price rise required for the resale to cover the price paid and the costs incurred to acquire.
What would overturn the verdict: a foundation carrying the cost of use — avoided rent, charges, maintenance — would make it possible to compare buying with renting, and thus to give “losing money” a meaning that these thresholds do not have.
Read the chapter that demonstrates it
Cannot be settled
“A rate can be renegotiated; a signed price cannot.”
The detailed answer, and the measurement that underpins it, can be read in the full study.
What this verdict does not cover: no series in the foundation carries loan buy-backs, the indemnities actually stipulated, or the costs of a renegotiation. The claim is not judged false: it is not testable here.
See why— 2 conclusions of the study
Why this idea circulates: the two halves of the sentence are accurate taken separately, and one of them is even a legal truism: the price entered in the deed is not reopened. From this it is concluded that the rate would be the repairable variable, hence the least serious to bear.
- relevant, without settling · value withheld
  The study quantifies the interest paid over eight years in two environments built at identical monthly burden.
- relevant, without settling · value withheld
  The study quantifies, in the two environments built at identical burden, what each exposes: the price fall tolerated by the resale at five years, and the debt remaining after eight years.
What would overturn the verdict: a public statistic on renegotiations and buy-backs — volume, rate gap obtained, total cost of the operation — would make the comparison possible.
Read the chapter that demonstrates it
Misleading
“It is always better to buy expensive with a very low rate.”
The detailed answer, and the measurement that underpins it, can be read in the full study.
What this verdict does not cover: the two environments are a CONSTRUCTION of the study: two rates actually observed, matched at equal monthly burden. No market offered these two deals to the same household on the same day, and the study does not say that a lower entry price is actually obtained when rates rise.
See why— 3 conclusions of the study
Why this idea circulates: a very low rate means paying little interest, which is true and measurable. The conclusion “so it is always better” silently adds that the entry price would not matter. The opposite claim — it is always better to buy cheaper with a high rate — circulates just as much, and runs into the same study.
- supports it · value withheld
  The study quantifies the interest paid over eight years in two environments built at identical monthly burden.
- goes against it · value withheld
  The study quantifies, in the two environments built at identical burden, what each exposes: the price fall tolerated by the resale at five years, and the debt remaining after eight years.
- relevant, without settling · value withheld
  The study quantifies, after eight years of amortisation, the price fall tolerated by the cash threshold and by the deposit-recovery threshold.
What would overturn the verdict: nothing in the available foundation: settling it would require attaching a probability to a price movement, which the contract of this study explicitly forbids writing.
Read the chapter that demonstrates it
Cannot be settled
“Notary fees are simply seven to eight per cent.”
The detailed answer, and the measurement that underpins it, can be read in the full study.
What this verdict does not cover: the share of acquisition costs is a PARAMETER declared by the study, not a measurement: it depends on the département, the price and the structure of the deed, and shifts by more than a point from one case to another. A gap of a few tenths of a point does not survive this uncertainty, and the study refuses to present it as a refutation.
See why— 2 conclusions of the study
Why this idea circulates: the order of magnitude is right for an existing home, and it is repeated everywhere. The word that causes the problem is “simply”: it suggests that everything an acquisition costs on top of the price fits in this single line.
- relevant, without settling · value withheld
  The study quantifies the price rise required for the resale to cover the price paid and the costs incurred to acquire.
- relevant, without settling · value withheld
  The study compares, across three deposit scenarios, the two thresholds that move with the holding period — and shows that they do not move in the same direction.
What would overturn the verdict: a public series of transfer duties actually paid per transaction, with their dispersion, would replace the parameter with a measurement.
Read the chapter that demonstrates it
Rather true
“A fall in rates automatically revives buying capacity.”
The detailed answer, and the measurement that underpins it, can be read in the full study.
What this verdict does not cover: BORROWING capacity and BUYING capacity are two distinct quantities. The price index used is constant-quality and has no monetary unit: the study forbids itself to multiply one by the other.
See why— 3 conclusions of the study
Why this idea circulates: the first half of the mechanism is exact and immediate: at equal monthly payment, a lower rate finances a larger capital. The second half — that BUYING power follows — cannot be deduced from it.
- supports it · value withheld
  The study quantifies what one extra point of rate removes from financeable capital over twenty years, step by step, and shows why the effect is not constant.
- supports it · value withheld
  The study quantifies what the same monthly effort has lost in financeable capital since the rate trough.
- relevant, without settling · value withheld
  The study measures by how much the existing-home price index moved from 2000 to 2025, as an annual average.
What would overturn the verdict: a price in euros per property, or per square metre, at a fine geographical level — which the constant-quality index does not provide and is not meant to provide.
Read the chapter that demonstrates it
False
“Waiting twelve months costs nothing if prices fall.”
The detailed answer, and the measurement that underpins it, can be read in the full study.
What this verdict does not cover: this cost is not complete: the household that waits does not pay its instalments either, and pays rent, which the foundation does not carry. The study serves the two quantities separately and never adds them up.
See why— 2 conclusions of the study
Why this idea circulates: the reasoning retains only one term, the visible one: the advertised price. What happens during the wait appears on no advertisement.
- goes against it · value withheld
  The study quantifies what a waiting period costs in capital that the household will not have repaid, for four delays and two loan terms.
- relevant, without settling · value withheld
  The study quantifies by how much the “monthly payment × term” multiplication overstates the capital actually borrowable, at identical monthly burden.
What would overturn the verdict: a rent series comparable to the scope of the study would make it possible to close the balance of waiting, which remains partial here and is declared as such.
Read the chapter that demonstrates it
Cannot be settled
“Fees borne by the buyer are always impossible to finance.”
The detailed answer, and the measurement that underpins it, can be read in the full study.
What this verdict does not cover: no series in the foundation carries financing decisions: neither the share of files in which separate fees are financed, nor the conditions that prevail. The study bounds the consequences of each hypothesis; it does not say which one occurs, and it does not turn a rule specific to the zero-interest loan into a general rule.
See why— 2 conclusions of the study
Why this idea circulates: a financing plan often runs into the fact that acquisition costs, for their part, are not financed — and this rule is extended to fees, which appear at the same moment and on the same deed. The confusion is all the easier because a refusal, when it comes, is experienced as a general rule rather than as an institution's decision.
- relevant, without settling · value withheld
  The study quantifies the additional cash that an arrangement borne by the buyer requires when the fees are not financed.
- relevant, without settling · value withheld
  The study quantifies what an arrangement in which the fees are borne by the buyer saves in acquisition costs, for an economically identical operation.
What would overturn the verdict: a credit-production statistic distinguishing what is actually released — price, fees, costs — would make the claim testable instead of plausible.
Read the chapter that demonstrates it
Misleading
“Fees borne by the seller are always more advantageous for the buyer.”
The detailed answer, and the measurement that underpins it, can be read in the full study.
What this verdict does not cover: the study compares a SCENARIO: fees set as a percentage, a lender whose decision is a hypothesis, and an economically identical operation on both sides. It observes neither the frequency of the arrangements, nor the price actually obtained at the end of a negotiation, nor what a buyer negotiates differently depending on whether the commission is displayed separately or not.
See why— 2 conclusions of the study
Why this idea circulates: the arrangement is simpler to read — one price, one negotiation — and it spares the buyer from having to find a separate sum. Both advantages are real; from them an “always” is drawn that forgets that transfer duties are then calculated on a higher base.
- goes against it · value withheld
  The study quantifies what an arrangement in which the fees are borne by the buyer saves in acquisition costs, for an economically identical operation.
- supports it · value withheld
  The study quantifies the additional cash that an arrangement borne by the buyer requires when the fees are not financed.
What would overturn the verdict: a matched observation of comparable transactions under the two arrangements — signed price, fees, financing obtained — would make it possible to leave the scenario and measure the real gap.
Read the chapter that demonstrates it
Misleading
“A bigger deposit is always preferable.”
The detailed answer, and the measurement that underpins it, can be read in the full study.
What this verdict does not cover: the study compares the deposit with no alternative investment: it carries neither a return, nor taxation, nor a savings horizon. “Preferable” in the full wealth sense is not what it settles.
See why— 3 conclusions of the study
Why this idea circulates: a higher deposit reduces the capital borrowed, hence the interest, and reassures the lending institution. All of this is exact — and unrelated to what happens at resale.
- goes against it · value withheld
  The study compares, across three deposit scenarios, the two thresholds that move with the holding period — and shows that they do not move in the same direction.
- supports it · value withheld
  The study quantifies, after eight years of amortisation, the price fall tolerated by the cash threshold and by the deposit-recovery threshold.
- relevant, without settling · value withheld
  The study quantifies, for a household that resells in order to move, the margin it has to trade up, under five hypotheses of uniform market rise.
What would overturn the verdict: a reference return on the household's savings would make it possible to measure what the immobilised deposit does not earn — the missing term of the comparison.
Read the chapter that demonstrates it

Two claims that this study cannot test

“The debt ceiling absolutely forbids going above thirty-five per cent.”
the study does carry the flexibility margin that the standard leaves to institutions, but it observes NO actual credit production: neither the share of files granted above the ceiling, nor their profile. Without this, saying whether the prohibition is absolute in practice would be an opinion, not a result.
“Europe wants to abolish French fixed-rate mortgage lending.”
no economic series tests a regulatory intention. This claim falls under the analysis of texts under discussion, not of a data foundation — and this study has neither.

Back to the three situations

Reserved

The real constraints of financing

Why “35% of income” is not enough to describe what a bank agrees to lend.

Reserved · full analysis

From income to the maximum price of the home, step by step and for 3 situations — two of them share the same income and are separated by the deposit: what the debt ceiling removes, what insurance takes, and by how much the “monthly payment × term” multiplication overstates the capital actually borrowable, at identical burden — with the six financing constraints and the table that measures them over 5 rate levels

Back to the three situations

Reserved

The three exit thresholds

From what resale price does the sale cover the loan, the deposit, then the costs incurred to buy?

Reserved · full analysis

The three exit thresholds plotted year by year for 3 situations, on a common scale: the price below which selling requires putting in cash, the one below which the deposit is not recovered, and the one below which acquisition costs are not covered — with the exact values at the 4 frozen horizons

Reserved

Higher price with lower rate, or lower price with higher rate?

Two complete housing environments, built at identical monthly burden — and what really separates them.

Reserved · full analysis

Two complete housing environments, set back to back at identical monthly burden — higher price with lower rate, lower price with higher rate: entry price, deposit, capital borrowed, interest, remaining debt and tolerated price fall, each gap quantified in euros and in points

Reserved

Should one wait three, six, nine or twelve months?

What would have to change for waiting to improve the project? The answer is an equivalence, not a forecast.

Reserved · full analysis

A grid of 35 combinations of price and rate, served for 2 loan terms: the reader sets their hypothesis, the engine returns what the same monthly effort would reach. And, separately — because the two quantities do not add up —, what 4 waiting periods cost in unrepaid capital. With no forecast or probability whatsoever

Back to the three situations

Reserved

The second purchase

Does the rise in the value of the first property improve the household's position, if the next home has also gone up?

Reserved · full analysis

The second-purchase chain, link by link — sale price, outstanding capital, indemnity, sum available after the sale, new loan, acquisition and financing costs — over 5 rise hypotheses that the reader varies, with the trade-up margin in euros and as a percentage of the resale price. And the four questions that arise at the same moment: what the early-repayment clause of your contract really changes, what acquisition costs cover and why a second purchase is no longer protected by the first-time-buyer exception, what a fee arrangement borne by the seller or by the buyer changes in duties and cash under three financing hypotheses, and why a high capital gain on a main residence is not taxed

Open to all

What this study cannot measure

These absences are not oversights: they decide what the study has no right to conclude.

the APR (TAEG) of mortgage loans, that is, the rate that includes borrower insurance, arrangement fees, guarantee and brokerage
Forbids asserting that a borrower pays exactly the rate served here. All the study's monthly payments are EXCLUDING INSURANCE as long as the insurance parameter is not added, and it is declared, not observed
the actual cost of borrower insurance, which depends on age, health, the share insured and the contract
Forbids presenting the total burden as measured. The study CALCULATES it from a declared rate, and a reader whose insurance costs twice as much will read a result that does not concern them
any local dimension: the price index is national, and there is no municipal price or price per square metre here
Forbids concluding anything about a city, a neighbourhood or a property. The national and local trends may have diverged by a factor of two over the window

The 5 other quantities that the foundation does not carryeach with what its absence forbids concluding

the income of BUYING households, which is neither the average income nor the median wage — buyers are a particular subset of the population
Forbids saying what a typical household can buy. The study's profiles are declared CALCULATION SCENARIOS, not observed households
the actual distribution of personal deposits and its evolution, which the catalogue does not carry
Forbids asserting that the required deposit has risen or fallen. The study VARIES the deposit to show its effect; it does not observe its level
the rent that would have been paid without the purchase, the return on a deposit invested elsewhere, works, property tax, charges and maintenance
Forbids saying whether buying was better than renting, and calling S3 a profitability, a wealth neutrality or a no-regret threshold
the machine-readable history of the usury ceilings applicable to MORTGAGE loans — the vehicle open as a service carries only consumer credit
Forbids plotting a usury constraint year by year. The ceiling is treated as a dated legal parameter, with its reference, and never interpolated between two periods of effect
the refusal rate of credit applications and the actual lending criteria of institutions
Forbids asserting that a file complying with the debt-ratio standard obtains a loan. The standard is a bound, not a guarantee of approval

Open to all

Method and evidence

Everything that allows this study to be contradicted: where its figures come from, what it sets itself, what it verified with the producer, and the exact list of what it asserts.

The 6 series used, and their producerfor each: the concept measured, its role in the study, and what it does not measure

European Central Bank (ECB)
Basis of the calculation
Interest rate on NEW home loans granted to households, France, annual average of the twelve monthly rates — effective rate in the narrow sense
Does not measure: the APR (TAEG), which adds borrower insurance, arrangement fees, guarantee and brokerage, and is structurally higher; nor the average rate on the STOCK of outstanding loans, which is markedly lower; nor the rate offered to an individual borrower
INSEE
Basis of the calculation
Notaires-INSEE price index for EXISTING homes, metropolitan France, all dwellings, base 100 in the 2015 annual average — constant quality
Does not measure: a price in euros, nor a price per square metre, nor a local market: the dispersion between municipalities far exceeds the national trend. The index does not cover new housing
INSEE
Context
Notaires-INSEE price index for existing APARTMENTS, metropolitan France
Does not measure: the relative price level compared with houses: the two indices share a common base, which informs on their CHANGES and not on their levels
INSEE
Context
Notaires-INSEE price index for existing HOUSES, metropolitan France
Does not measure: the floor area acquired: a house and an apartment with the same index represent neither the same property nor the same budget
INSEE
Context
Median net wage in full-time equivalent in the private sector, France — the employee in the middle of the distribution
Does not measure: the income of a HOUSEHOLD, which may include two salaries, non-wage income and benefits. A buying household is not a median employee
INSEE
Context
Consumer price index excluding tobacco, all households, France — base 2025
Does not measure: the housing cost for a buyer-occupier: the index covers rents and charges, not loan repayments

The 9 parameters of the study, and their naturewhat a text fixes, and what the study sets itself

Maximum debt-service ratio35%Fixed by a text
Maximum loan term25 yearsFixed by a text
Flexibility margin left to institutions20% of quarterly productionFixed by a text
Early-repayment indemnity — ceiling in months of interest6 months of interest on the capital repaidFixed by a text
Early-repayment indemnity — ceiling as a share of the outstanding capital3% of the outstanding capitalFixed by a text
Acquisition costs for existing homes7.5% of the priceDeclared
Borrower insurance rate0.34% per year of the initial capitalDeclared
Guarantee and arrangement fees1.4% of the amount borrowedDeclared
Agency fees used for the scenario5% of the advertised priceDeclared

Comparison with the producer — the base yearwhat the producer asserts about its own method, and what the study obtains by redoing it

The index is titled “base 100 in the 2015 annual average”. This title is a statement by the producer about its own method: the average of the four quarters of 2015 must equal one hundred. The study moves from quarter to year by this same average.

Series	Calculated by the study	Asserted by the producer	Gap
All	100.000	100.000	+0.000
Apartments	100.025	100.000	+0.025
Houses	99.975	100.000	−0.025

The overall index lands exactly on 100.000. The two sub-series deviate from it by twenty-five thousandths of a point. As quarterly values are published to one decimal, each carries an error of at most five hundredths, and so does their average: the gap observed is therefore HALF of what simple rounding allows. It is a publication residual, not a methodological divergence.

The exact formula, and the sentences the study forbids itself to writewhat is needed to redo the calculation, and the list of what it will never conclude

Financeable capital

K = m × (1 − (1 + i)^−n) ÷ i, where m is the monthly payment excluding insurance, i the annual rate divided by twelve, and n the number of instalments. The total burden adds the insurance premium, calculated on the initial capital.

The operation this study never performs

Multiplying a financeable capital by a change in a price index to derive “the home you could have bought”. The index is constant-quality and has no monetary unit: the product of the two designates no real property.

The 8 sentences the study forbids itself to write

never call S3 a profitability, a wealth neutrality or a no-regret threshold — it ignores the alternative cost of use, so it cannot compare any option with any other
never write what the market WILL do, nor attach a probability to a scenario: scenarios are conditional, and their condition is written before their result
never recommend buying, selling, waiting or negotiating — the study shows a mechanism, it prescribes no course of action
never write that the fall in rates CAUSED the rise in prices: the two moved together, and no method used here identifies a causal effect
never present a result as the reader's borrowing capacity: no data here observes their situation, and the debt-ratio standard is a bound, not a promise of approval
never turn a financeable capital into a “home you could have bought”: that would require a price in euros per property, which the index does not carry
never apply the 2021 debt-ratio standard to the years that precede it — it did not exist, and lending practices at the time were different
never present a monthly payment calculated at the MIR rate as the total burden: this rate carries interest alone, and insurance is a declared parameter

Everything this study asserts — the 20 frozen conclusionseach with its nature, its period, its scope, its source, its formula and its limits

The study measures by how much the existing-home price index moved from 2000 to 2025, as an annual average.
FactConfidence: HighReserved value
Value
reserved
Unit
index change, base 100 in the 2015 annual average
Period
2000-2025
Scope
Metropolitan France — existing homes, main residence
Formula
(index of the last year ÷ index of the opening year − 1) × 100
Revision
the latest published quarters are provisional and may be revised by the producer
Sources
INSEE — Notaires-INSEE price indices for existing homes, series 010567058
- Limit: the index is national: the dispersion between municipalities far exceeds the national trend
- Limit: the index is constant-quality and has no monetary unit: it cannot be converted into the price of a property
The peak of the index, as an annual average, is reached in 2022; from 2022 to 2025, it fell by 4.2%.
FactConfidence: High
Value
−4.2%
Unit
index change
Period
2022-2025
Scope
Metropolitan France — existing homes, main residence
Formula
(index of the last year ÷ index of the peak year − 1) × 100
Revision
revisable — the peak may move if a recent year is revised
Sources
INSEE — Notaires-INSEE price indices for existing homes, series 010567058
- Limit: the index is national: the dispersion between municipalities far exceeds the national trend
- Limit: a peak identified on annual averages is not a quarterly peak
The study dates the lowest point of the rate on new home loans over the 2000–2025 window, and gives its value.
FactConfidence: HighReserved value
Value
reserved
Unit
% per year
Period
2000-2025
Scope
France — new home loans granted to households
Formula
minimum of the annual average of the twelve monthly rates
Revision
final for closed years
Sources
ECB — MIR statistics (MIR.M.FR.B.A2C.A.R.A.2250.EUR.N), supplied by the Banque de France
- Limit: the rate served carries interest alone: it is neither the APR (TAEG) nor the total burden
- Limit: national average: no individual borrower obtained this exact rate
The average of the four raw quarters of 2015 is 100.000 — exactly the base declared by the producer.
FactConfidence: High
Value
100.000
Unit
index
Period
2015
Scope
Metropolitan France — existing homes, main residence
Formula
arithmetic mean of the four raw quarterly values of the base year
Revision
the base year is fixed by the producer and is not revised
Sources
INSEE — Notaires-INSEE price indices for existing homes, series 010567058
- Limit: the two sub-series deviate by twenty-five thousandths of a point, i.e. half of what rounding the published values to one decimal allows
The study quantifies what one extra point of rate removes from financeable capital over twenty years, step by step, and shows why the effect is not constant.
Derived resultConfidence: HighReserved value
Value
reserved
Unit
change in financeable capital
Period
not applicable — actuarial relationship
Scope
amortising loan with constant annuities, fixed rate, excluding insurance
Formula
K = m × (1 − (1 + i)^−n) ÷ i, compared between two rates one point apart
Revision
not applicable — mathematical relationship, not revisable
Sources
study engine — inverse annuity
- Limit: the rate served carries interest alone: it is neither the APR (TAEG) nor the total burden
- Limit: the effect is not constant: it decreases as the starting level of rates rises
The study quantifies what the same monthly effort allowed one to borrow in each year from 2000 to 2025, and the gap between the best and the worst year.
Derived resultConfidence: HighReserved value
Value
reserved
Unit
gap between the maximum and the minimum of the window
Period
2000-2025
Scope
€1,000 a month, twenty years, observed average annual rate
Formula
inverse annuity applied to the average rate of each year, at fixed monthly payment and term
Revision
follows any revisions of the rate series
Sources
ECB — MIR statistics (MIR.M.FR.B.A2C.A.R.A.2250.EUR.N), supplied by the Banque de France · study engine — inverse annuity
- Limit: BORROWABLE capital, never an affordable home
- Limit: no lending standard is applied to years before its entry into force
The study quantifies what the same monthly effort has lost in financeable capital since the rate trough.
Derived resultConfidence: HighReserved value
Value
reserved
Unit
change in financeable capital
Period
2021-2025
Scope
€1,000 a month, twenty years
Formula
inverse annuity at the trough rate, compared with the inverse annuity at the rate of the last closed year
Revision
follows any revisions of the rate series
Sources
ECB — MIR statistics (MIR.M.FR.B.A2C.A.R.A.2250.EUR.N), supplied by the Banque de France · study engine
- Limit: the rate served carries interest alone: it is neither the APR (TAEG) nor the total burden
- Limit: this change says nothing about prices: comparing it with the change in the index would be another calculation
The gap between the apartments index and the houses index reaches 14.2 points in 2007.
Derived resultConfidence: High
Value
−14.2 points in 2007
Unit
index points, common base
Period
1996-2025
Scope
Metropolitan France — existing homes, main residence
Formula
difference of the two indices, year by year, on the common base
Revision
follows the producer's revisions
Sources
INSEE — series 010567056 and 010567060
- Limit: two indices sharing a base inform on their CHANGES, never on their relative levels
- Limit: the geographical composition of the two markets differs
The study quantifies by how much applying the debt ceiling to the repayment instalment alone overstates financeable capital, at five rate levels.
Result under a declared parameterConfidence: MediumReserved value
Value
reserved
Unit
overstatement of financeable capital
Period
not applicable — actuarial relationship
Scope
declared insurance rate of 0.34% a year of the initial capital
Formula
capital under total burden = burden ÷ (unit annuity + monthly insurance rate), compared with the capital calculated without the second term
Revision
not applicable — depends on a declared parameter, not on a publication
Sources
study engine · declared parameter of the study
- Limit: the result depends on parameters set by the study, which vary from one case to another and are not measured
- Limit: a borrower whose insurance costs twice as much reads a gap twice as large
The study quantifies the price rise required for the resale to cover the price paid and the costs incurred to acquire.
Result under a declared parameterConfidence: MediumReserved value
Value
reserved
Unit
price change required
Period
not applicable — holding horizon
Scope
deposit of 20% of the price, acquisition costs declared at 7.5%
Formula
threshold = purchase price + transfer duties, notary's fees, guarantee and arrangement fees, as a ratio to the purchase price
Revision
not applicable — depends on declared parameters
Sources
study engine · declared parameters of the study
- Limit: the result depends on parameters set by the study, which vary from one case to another and are not measured
- Limit: this threshold measures neither profitability, nor wealth neutrality, nor regret: it ignores the alternative cost of use
- Limit: agency fees at resale are not included
The study quantifies, after eight years of amortisation, the price fall tolerated by the cash threshold and by the deposit-recovery threshold.
Result under a declared parameterConfidence: MediumReserved value
Value
reserved
Unit
price fall tolerated
Period
not applicable — holding horizon
Scope
deposit of 20% of the price, acquisition costs declared at 7.5%
Formula
cash threshold = outstanding capital + indemnity; recovery threshold = cash threshold + deposit; both as a ratio to the purchase price
Revision
not applicable — depends on declared parameters
Sources
study engine · declared parameters of the study
- Limit: the result depends on parameters set by the study, which vary from one case to another and are not measured
- Limit: amortisation assumes a loan run with no partial repayment
The study quantifies the interest paid over eight years in two environments built at identical monthly burden.
Result under a declared parameterConfidence: MediumReserved value
Value
reserved
Unit
cumulative interest at eight years
Period
not applicable — two environments built on observed rates
Scope
same monthly burden, same term, same deposit share; only the rate changes
Formula
constant-annuity repayment schedule run on each environment, cumulative interest at ninety-six instalments
Revision
not applicable — depends on declared parameters
Sources
ECB — MIR statistics (MIR.M.FR.B.A2C.A.R.A.2250.EUR.N), supplied by the Banque de France · study engine · declared parameters of the study
- Limit: the result depends on parameters set by the study, which vary from one case to another and are not measured
- Limit: the two rates are observed in the series but their pairing into “environments” is a construction of the study
The study quantifies, for a household that resells in order to move, the margin it has to trade up, under five hypotheses of uniform market rise.
Result under a declared parameterConfidence: MediumReserved value
Value
reserved
Unit
euros of trade-up margin
Period
not applicable — resale horizon of eight years
Scope
the first-time-buyer couple of situation B, eight years after its first purchase; income and burden unchanged between the two purchases
Formula
sum available after the sale = sale price − outstanding capital − indemnity; maximum price of the new home = (sum available + loan net of guarantee) ÷ (1 + acquisition costs); trade-up margin = maximum price − sale price
Revision
not applicable — depends on declared parameters
Sources
study engine · declared parameters of the study
- Limit: the result depends on parameters set by the study, which vary from one case to another and are not measured
- Limit: the rise is assumed UNIFORM across the two properties: a market in which the targeted segment rises more slowly would give the opposite result
- Limit: income and credit conditions are held constant between the two purchases
- Limit: no agency fee is deducted from the sale price: the chain assumes a sale without an intermediary
The study quantifies what an arrangement in which the fees are borne by the buyer saves in acquisition costs, for an economically identical operation.
Result under a declared parameterConfidence: MediumReserved value
Value
reserved
Unit
euros of acquisition costs saved
Period
not applicable — transaction scenario
Scope
same property, same net sum to the seller, same remuneration of the professional; fees distinguished in the deed and mandate designating the buyer as debtor
Formula
saving = fees × acquisition-cost share, the duties being calculated on the seller's net price instead of the price including fees
Revision
not applicable — depends on declared parameters
Sources
study engine · declared parameters of the study · BOFiP BOI-ENR-DMTOI-10-10-20-20
- Limit: the result depends on parameters set by the study, which vary from one case to another and are not measured
- Limit: the fees used are an ASSUMPTION: they are free, no public statistic observes their distribution, and the study fabricates no average of them
- Limit: exclusion from the base requires TWO cumulative conditions — a mandate designating the buyer as debtor, and the amount distinguished in the deed
The study quantifies the additional cash that an arrangement borne by the buyer requires when the fees are not financed.
Result under a declared parameterConfidence: MediumReserved value
Value
reserved
Unit
euros of additional cash
Period
not applicable — transaction scenario
Scope
assumption in which the lender finances only the price of the property; acquisition costs are paid from own funds in both arrangements
Formula
cash = total cost of the operation − amount financed by the lender; the gap between the two arrangements equals the unfinanced fees less the saving in acquisition costs
Revision
not applicable — depends on declared parameters
Sources
study engine · declared parameters of the study
- Limit: the result depends on parameters set by the study, which vary from one case to another and are not measured
- Limit: the study observes NO bank decision: the three financing hypotheses bound what can happen, they measure no frequency
- Limit: the rules of the zero-interest loan expressly include negotiation fees borne by the buyer in the total cost of the operation; this does not hold for other loans
- Limit: this cash requirement is a FLOOR, not an impossibility: this household's deposit covers it comfortably. The constraint bites only on a household whose deposit does not exceed this floor
The study quantifies the share of insurance in the monthly burden, at five rate levels.
Result under a declared parameterConfidence: MediumReserved value
Value
reserved
Unit
share of the monthly burden
Period
not applicable — actuarial relationship
Scope
declared insurance rate of 0.34% a year of the initial capital
Formula
monthly premium ÷ (repayment instalment + monthly premium)
Revision
not applicable — depends on a declared parameter
Sources
declared parameter of the study · study engine
- Limit: the result depends on parameters set by the study, which vary from one case to another and are not measured
- Limit: the convention used applies the rate to the INITIAL capital; a delegated contract often applies it to the outstanding capital, which costs less
The study quantifies by how much the “monthly payment × term” multiplication overstates the capital actually borrowable, at identical monthly burden.
Result under a declared parameterConfidence: MediumReserved value
Value
reserved
Unit
overstatement of borrowable capital
Period
not applicable — actuarial relationship
Scope
average rate of 2025, term of 20 years, insurance declared at 0.34% a year
Formula
(monthly burden used × 12 × term) relative to the capital under total burden = same burden ÷ (unit annuity + unit premium); both terms start from the SAME burden, failing which the gap would mix the effect of the rate with a change of denominator
Revision
not applicable — depends on a declared parameter and on the rate used
Sources
study engine · declared parameter of the study
- Limit: the result depends on parameters set by the study, which vary from one case to another and are not measured
- Limit: the naive multiplication is not a banking practice: it is the reading shortcut that the chapter refutes
- Limit: the gap does NOT contain the move from the debt ceiling to the burden used by the study: that is shown separately, as a rung of the cascade
- Limit: the invariance across profiles holds because the burden is proportional to income; it would cease if the debt-service ratio used varied from one profile to another
The study compares, across three deposit scenarios, the two thresholds that move with the holding period — and shows that they do not move in the same direction.
Result under a declared parameterConfidence: MediumReserved value
Value
reserved
Unit
price fall tolerated at eight years
Period
not applicable — holding horizon
Scope
deposits of 10% and 35% of the price, same term, same rate
Formula
exit thresholds calculated on three operations built from the same monthly burden and different deposit shares
Revision
not applicable — depends on declared parameters
Sources
study engine · declared parameters of the study
- Limit: the result depends on parameters set by the study, which vary from one case to another and are not measured
- Limit: the two thresholds compared do not measure the same thing: one says whether the sale is possible, the other whether it returns to the household what it had paid in
- Limit: a higher deposit is not compared with the return the same sum would have produced elsewhere: the study carries no alternative investment
The study quantifies, in the two environments built at identical burden, what each exposes: the price fall tolerated by the resale at five years, and the debt remaining after eight years.
Result under a declared parameterConfidence: MediumReserved value
Value
reserved
Unit
price fall tolerated, and outstanding capital
Period
not applicable — holding horizons
Scope
same monthly burden, same term, same deposit share; only the rate changes
Formula
cash threshold = (outstanding capital + indemnity) as a ratio to the purchase price; outstanding capital by closed form at ninety-six instalments
Revision
not applicable — depends on declared parameters
Sources
ECB — MIR statistics (MIR.M.FR.B.A2C.A.R.A.2250.EUR.N), supplied by the Banque de France · study engine · declared parameters of the study
- Limit: the result depends on parameters set by the study, which vary from one case to another and are not measured
- Limit: the threshold is expressed as a PERCENTAGE of the purchase price: as the two purchase prices differ, the same percentage fall does not represent the same sum
- Limit: carrying less debt is not an advantage in itself: the lower debt goes with a cheaper property, and the study establishes no preference between the two
The study quantifies what a waiting period costs in capital that the household will not have repaid, for four delays and two loan terms.
Result under a declared parameterConfidence: MediumReserved value
Value
reserved
Unit
capital not repaid during the delay
Period
not applicable — waiting period
Scope
monthly burden of the two-median-income scenario, average rate of 2025, loan of 20 years
Formula
capital borrowed − outstanding capital after the number of months of waiting, on the repayment schedule of the operation the household can finance at the average rate of the last closed year
Revision
not applicable — depends on declared parameters
Sources
study engine · declared parameters of the study
- Limit: the result depends on parameters set by the study, which vary from one case to another and are not measured
- Limit: this amount is NOT the full cost of waiting: the household that waits does not pay the instalments either, and pays rent, which the study does not carry
- Limit: it is compared with no price change: the two quantities are served separately, and the study does not add them up

The 4 reading aidsthe words without which a sentence of this study changes meaning

Annuity: the monthly payment of a loan with equal instalments: it repays mostly interest at first, then mostly capital, in a proportion that changes with each instalment.; This is not: this is not the total burden: borrower insurance is added to it and is not included in the displayed rate.
Outstanding capital: what remains to be repaid to the bank at a given moment, excluding future interest. It is the sum that the sale must cover first.; This is not: this is not what the home is worth, nor what the household has already paid — interest paid is not part of it.
Debt-service ratio: the share of income absorbed by the credit burden, insurance included. The lending standard caps it, with a flexibility margin left to institutions.; This is not: this is not a right to borrow: staying below the ceiling guarantees no approval, and exceeding it is not always refused.
Constant-quality index: a price index built to neutralise the fact that the homes sold in one year are not those sold the next — without which a change in composition would pass for a price movement.; This is not: this is neither a price nor a price per square metre: the index has no monetary unit and cannot be converted into euros.

Was it necessary to separate houses from apartments?a method decision, taken before looking at the data

Yes. The gap between the two indices reaches 14.2 points in 2007, above the 5-point threshold set before the calculation.

The threshold was set before looking at the data: five index points on a common base. The gap of 2007 stands at −14.2 percentage points; at the end of the series it is only −1.5 percentage points. The two markets therefore drifted apart, then came closer — which the overall curve alone did not show.

Cite this studythe reference to copy, the edition, the version and the data cut-off date

France ÉcoScan, “Can you still become a homeowner in France?”, edition 2026.2 — main residence in the existing market, France ÉcoScan v9.0.14.

Edition: 2026.2
Product version: France ÉcoScan v9.0.14
Published on: 2026-09-15
Data cut-off: 2026-09-15
Method: financeable capital by inverse annuity at a proportional nominal rate; exit thresholds by flow accounting on the repayment schedule; independent reproduction by a second route with no closed form.

The contents

What this study contains

Open to all

The answer to the question in one sentence, the principle of the mechanism in one line, and what the study does not settle
Three dated benchmarks from the latest published reading: the rate for the latest month, the price of existing homes over one year at the latest quarter, and what the same monthly payment buys over one year
1 interactive chart as an example — the price of existing homes by type of dwelling since 1996, with the reading of the selected year
The 8 quantities that the foundation does not carry — including the local price, the income of buying households and the actual cost of insurance — and what their absence forbids concluding
The 3 situations that the study follows from end to end — income used, credit burden and deposit share of each, with what none of them claims to be
9 common beliefs tested against the frozen conclusions of the study, each with a verdict CALCULATED from its evidence and not written by hand — the detailed answer remaining in the full study — and 2 claims set aside for lack of a foundation to test it
The 6 series used, each with its producer, its official identifier, its scope and what it does not measure
The comparison with the producer on 3 series: the annual average of the base year must equal one hundred, and the study shows what it obtains
The exact formula for financeable capital, written to be copied, and the 8 sentences the study forbids itself to write
The 20 quantified conclusions of the study, each with its nature — fact, derived result or result under a declared parameter —, its period, its scope, its source, its formula and its limits; the value of the reserved conclusions remains in the full study
Reading aids: annuity, outstanding capital, debt-service ratio, constant-quality index, defined on the page
The reference to copy, the version, the edition and the data cut-off date — to cite this study in a note

Reserved

The quantified, dated answer, over 26 years of rates and 30 years of prices, extended to the latest published rate month — and the state of the last closed year, benchmark by benchmark
What one point of rate really removes from financeable capital, over 2 terms and five rate levels, step by step — with the reason why the effect is not constant
What the same monthly effort allowed one to borrow depending on the period, year by year from 2000 to 2025, at the rate actually observed — the most favourable year, the least favourable and the last closed year
From income to the maximum price of the home, step by step and for 3 situations — two of them share the same income and are separated by the deposit: what the debt ceiling removes, what insurance takes, and by how much the “monthly payment × term” multiplication overstates the capital actually borrowable, at identical burden — with the six financing constraints and the table that measures them over 5 rate levels
The three exit thresholds plotted year by year for 3 situations, on a common scale: the price below which selling requires putting in cash, the one below which the deposit is not recovered, and the one below which acquisition costs are not covered — with the exact values at the 4 frozen horizons
Two complete housing environments, set back to back at identical monthly burden — higher price with lower rate, lower price with higher rate: entry price, deposit, capital borrowed, interest, remaining debt and tolerated price fall, each gap quantified in euros and in points
2 premium charts — financeable capital against the rate from 2000 to 2025, and outstanding capital in the two environments, year by year, with what to see in them and what they do not mean
A grid of 35 combinations of price and rate, served for 2 loan terms: the reader sets their hypothesis, the engine returns what the same monthly effort would reach. And, separately — because the two quantities do not add up —, what 4 waiting periods cost in unrepaid capital. With no forecast or probability whatsoever
The second-purchase chain, link by link — sale price, outstanding capital, indemnity, sum available after the sale, new loan, acquisition and financing costs — over 5 rise hypotheses that the reader varies, with the trade-up margin in euros and as a percentage of the resale price. And the four questions that arise at the same moment: what the early-repayment clause of your contract really changes, what acquisition costs cover and why a second purchase is no longer protected by the first-time-buyer exception, what a fee arrangement borne by the seller or by the buyer changes in duties and cash under three financing hypotheses, and why a high capital gain on a main residence is not taxed